US 7398209

Here's a concise summary of US Patent 7398209:

US Patent 7398209: Systems and methods for responding to natural language speech utterance

• Title: Systems and methods for responding to natural language speech utterance
• Assignee: Dialect LLC (Previously VoiceBox Technologies Corp, then VoiceBox Technologies, Inc., then Nuance Communications, Inc., then Voice Inventions, LLC)
• Inventors: Robert A. Kennewick, David Locke, Michael R. Kennewick, SR., Michael R. Kennewick, JR., Richard Kennewick, Tom Freeman
• Filing Date: June 3, 2003
• Issue Date: July 8, 2008
• Abstract: The patent describes systems and methods for receiving and executing natural language queries and/or commands. It aims to overcome limitations of prior speech query and response systems by providing a complete speech-based information query, retrieval, presentation, and command environment. This environment significantly leverages context, prior information, domain knowledge, and user-specific profile data to create a natural interaction for users across multiple domains. The system creates, stores, and uses extensive personal profile information for each user to improve the reliability of determining context and presenting relevant results for questions or commands.

Plain-Language Overview of Independent Claims:

The patent has several independent claims (Claim 1, Claim 15, Claim 16, Claim 24, Claim 32, Claim 33, Claim 41, Claim 49, Claim 57, Claim 65, Claim 73, Claim 81, Claim 89, Claim 97, Claim 105, Claim 113, Claim 121, Claim 129, Claim 137, Claim 145, Claim 153, Claim 161, Claim 169, Claim 177, Claim 185, Claim 193, Claim 201, Claim 209). Here's an overview of the first few and representative independent claims:

• Claim 1: This claim describes a computer-implemented method for responding to natural language speech. It involves capturing a user's speech utterance, interpreting its meaning, determining the required domain of expertise and context, formulating one or more machine-processable queries or commands, executing them, evaluating and combining the results, and then presenting a natural language speech response to the user. The method emphasizes using context, prior information, domain knowledge, and user profile data, and handling partial failures with probabilistic or fuzzy reasoning.
• Claim 15: This claim describes a system that performs the method of Claim 1. It includes a speech unit interface (e.g., with a microphone) to receive speech, and a computer system with a speech recognition module, a parser, an event manager, a dictionary, user profile module, personality module, agent module, update manager, and databases. The event manager mediates interactions, and agents handle domain-specific functionality.
• Claim 16: This claim focuses on a method for processing a natural language speech command to control devices. It involves capturing the command, interpreting it, determining the command's domain and context, formulating device-specific commands, routing them to the relevant system or external devices, receiving and processing results (including errors), and optionally providing a response to the user about the command's success or failure.
• Claim 24: This claim outlines a computer program product stored on a computer readable medium for responding to natural language speech utterances. The program performs the steps of interpreting an utterance to determine its meaning, determining the domain and context, formulating machine-processable queries or commands, executing them, evaluating results, and presenting a response.
• Claim 32: This claim describes a system with a speech unit interface and a computer system, similar to Claim 15, but specifically for executing local or remote control functions based on natural language speech commands. It includes a network interface for communicating with external devices.
• Claim 33: This claim details a method for updating a natural language speech system. It involves periodically querying a source for licensed agents and database components, downloading and installing updates (executables, scripts, or data) as they become available, and uninstalling agents no longer in use.
• Claim 41: This claim describes a computer program product for updating a natural language speech system, performing the steps outlined in Claim 33.
• Claim 49: This claim describes a method for presenting information in response to a natural language speech query. It includes extracting relevant information from query results using probabilistic or fuzzy set matching, presenting these subsets first, and providing user commands to browse or manage the presented information (e.g., skip, find keywords, stop).
• Claim 57: This claim covers a computer program product that implements the method of Claim 49.
• Claim 65: This claim describes a method for personalizing a natural language speech system. It involves creating, storing, and using personal profile information for each user, and automatically updating this information as the user interacts with the system. The profile data is used in interpreting questions, formulating queries, interpreting results, and presenting answers.
• Claim 73: This claim covers a computer program product that performs the personalization method of Claim 65.
• Claim 81: This claim describes a method for resolving ambiguity in a user's natural language speech utterance. It involves initially interpreting the utterance, scoring the interpretation, and if the confidence is unsatisfactory, requesting the user to verify the interpretation and rephrase or provide additional information.
• Claim 89: This claim covers a computer program product that implements the ambiguity resolution method of Claim 81.
• Claim 97: This claim describes a method for filtering speech input in a natural language speech system. It involves using an array microphone to receive speech, adjusting its beam pattern to maximize gain toward the user and null noise, and processing the speech with analog or digital filters to optimize signal-to-noise ratio, cancel echoes, and notch out narrow-band noise.
• Claim 105: This claim covers a computer program product that implements the speech filtering method of Claim 97.
• Claim 113: This claim describes a method for handling multiple users in a natural language speech system. It involves recognizing each user (by name or voice), invoking their correct profile, and applying the correct profile and context for each utterance, even in overlapping or interleaved sessions. For security, it includes user verification (e.g., voiceprint matching).
• Claim 121: This claim covers a computer program product that implements the multi-user handling method of Claim 113.
• Claim 129: This claim describes a method for enabling content providers and users to extend a natural language speech system. This includes creating new agents, scripting existing agents, adding data to agents or databases, and adding or modifying links to information sources.
• Claim 137: This claim covers a computer program product that implements the system extension method of Claim 129.
• Claim 145: This claim describes a method for using dynamically invokeable personalities in a natural language speech system. It involves creating personalities with simulated behavioral characteristics (e.g., sympathy, irritation), randomizing aspects of responses (terms used, presentation order), and invoking personality characteristics using probabilistic or fuzzy set methods based on context, user interaction history, and preferences.
• Claim 153: This claim covers a computer program product that implements the dynamically invokeable personalities method of Claim 145.
• Claim 161: This claim describes a method for adapting speech encoding for transmission over a bandwidth-limited wireless link. It involves digitizing speech using adaptive lossy audio compression to optimize bandwidth while preserving components for optimal recognition, buffering the digitized speech, and adapting the output data rate to available bandwidth.
• Claim 169: This claim covers a computer program product that implements the adaptive speech encoding method of Claim 161.
• Claim 177: This claim describes a method for generating natural language speech messages without user prompting, such as alert messages.
• Claim 185: This claim covers a computer program product that implements the method of Claim 177.
• Claim 193: This claim describes a method for allowing a user to specify the spelling of words to improve speech recognition accuracy, using names of letters or a phonetic alphabet.
• Claim 201: This claim covers a computer program product that implements the spelling specification method of Claim 193.
• Claim 209: This claim describes a method for dynamically updating probabilities or fuzzy possibilities for words in a speech recognition dictionary and phrase tables to maximize correct recognition. The updates are based on criteria like application domain, questions/commands, contexts, user profile, and dialog history.

CAFC 2026 Dockets:
As of April 26, 2026, a search of CAFC 2026 dockets did not return any specific litigation or precedential rulings directly citing US Patent 7398209. The provided search results mention CAFC activities in 2026 concerning other patent numbers and general patent litigation trends.